Implementation Of Land Cover Change Detection Based On Supervised Classifications Of Multispectral Satellite Data For Leveraging Internet Of Things
The study reported in this paper aims to detect land cover changes using multispectral and multitemporal remote sensing data. The data came from Landsat TM satellite covering the area of Klang, located in Selangor, Malaysia. Initially, pre-processing was carried out to identify the stability of thre...
Saved in:
| Main Authors: | , , , , , , , |
|---|---|
| Format: | Article |
| Language: | en |
| Published: |
Asian Research Publishing Network (ARPN)
2019
|
| Subjects: | |
| Online Access: | http://eprints.utem.edu.my/id/eprint/24068/2/jeas_0519_7753%20IOT.pdf http://eprints.utem.edu.my/id/eprint/24068/ http://www.arpnjournals.org/jeas/research_papers/rp_2019/jeas_0519_7753.pdf |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | The study reported in this paper aims to detect land cover changes using multispectral and multitemporal remote sensing data. The data came from Landsat TM satellite covering the area of Klang, located in Selangor, Malaysia. Initially, pre-processing was carried out to identify the stability of three supervised methods namely maximum likelihood (ML), neural network (NN) and support vector machines (SVM) as the size of training pixels changed For this purpose, Landsat bands 1, 2, 3, 4, 5 and 7 for the year 1998 were used as the input for each of these methods to classify land covers within the study area. The generated land cover classifications were evaluated by statistically comparing each land cover with a reference data set using a confusion matrix. Subsequently, these methods were used to classify land covers of the same area
using Landsat data acquired in the year 2000 and 2005. The 2005 classification was then statistically compared with the
2000 classification using a confusion matrix for each of the methods. This produced land cover changes that occurred
between 2000 and 2005 which were generated using SVM, ML and NN. Results showed that land cover change detection
using SVM was quantitatively and qualitatively more accurate compared to ML and NN mainly due to the least affected by the size of training pixels. The findings of the study are relevant and beneficial in leveraging the internet of things practices. |
|---|